Environmental Impact Assessment Review 68 (2018) 81–89

Contents lists available at ScienceDirect

Environmental Impact Assessment Review

journal homepage: www.elsevier.com/locate/eiar

Systemising gender integration with rural stakeholders' sustainability MARK

impact assessments: A case study with three low-input upgrading strategies
F. Graefa,⁎, L.E.A. Hernandezb, H.J. Königa, G. Uckertc, M.T. Mnimbod
a
Leibniz Centre for Agricultural Landscape Research (ZALF) e. V., Institute of Land Use Systems, 15374 Müncheberg, Germany
b
Humboldt University, Faculty of Agriculture and Horticulture, Unter den Linden 6, 10099 Berlin, Germany
c
Leibniz Centre for Agricultural Landscape Research (ZALF) e. V., Institute of Socio-Economics, 15374 Müncheberg, Germany
d
Sokoine University of Agriculture, Development Studies Institute, Morogoro, Tanzania

A R T I C L E I N F O A B S T R A C T

Keywords: Participatory action research across food value chains (FVC) can help stabilise the food security of subsistence
Gender, upgrading strategies farmers by implementing upgrading strategies (UPS). These strategies can be assessed ex-ante and ex-post for
Impact assessment their potential social, ecological and economic sustainability impacts.
Rural food systems UPS implementation, however, often entails gender-specific changes and challenges in a farmer's social life,
Food value chains
economy and environment that either were not perceived and anticipated beforehand or are not followed up
Tanzania
during UPS implementation. Before and during their implementation, therefore, UPS need to be entirely un-
Participatory research
Sustainability derstood and assessed by both genders in terms of their potential social, ecological and economic sustainability
FoPIA impacts.
This article conceptualises a systematic framework for integrating gender in sustainability impact assessments
and presents gender-based assessment differences in three low-input UPS in Tanzanian FVC. We conducted ex-
ante and ex-post impact assessments using nine food security criteria developed earlier by the authors following
the Framework of Participatory Impact Assessment (FoPIA). Sustainability impact assessments—to a greater
extent than expected—differed to various extents between the genders for a) different food security criteria, b)
different sustainability dimensions (economic, social, and environmental), c) different points in time (T0, T1) of
assessments, d) different implemented UPS, and e) different members within the groups of female and/or male
stakeholders.
The results demonstrate the substantial importance of integrating female-male segregated assessments and
perceptions before and while implementing food-securing UPS. We anticipate that integrating these assessments
and perceptions as regular components will lead to better gendered social learning for both scientists and sta-
keholders and a holistic understanding of complex local food systems.

1. Introduction inclusion while targeting specific socio-cultural, ecological and eco-

Most of Africa's agricultural commodities are produced by poor
smallholders (IAASTD, 2009). Several development initiatives have
focussed on enhancing agricultural production and overall food se-
curity. However, the Sahel food crisis of 2011 and 2012 demonstrated
the need for better integration of entire food systems and the devel-
opment of location-specific strategies. Research and development
(R & D) projects have often been top-down and researcher-oriented
while including only a few disciplines, thus limiting their success. More
recent R & D projects have encompassed entire food systems (CGIAR,
2012; Millennium Villages, 2013), food value chains (Graef et al., 2014;
Gomez et al., 2011), and local and regional participatory stakeholder
nomic environments (Grimble and Wellard, 1997; König et al., 2013).
To reduce negative impacts and enhance livelihoods, the possible
impacts of these R & D projects should be assessed both before (ex-ante)
and after (ex-post) their implementation (König et al., 2013; Morris
et al., 2011). This can be done using sustainability impact assessments
that channel decision-making towards sustainability (Bond and
Morrison-Saunders, 2011; Morrison-Saunders et al., 2014), while
methodologically linking food security issues and the sustainability of
upgrading strategies (UPS) with the local setting (Schindler et al.,
2016a, 2016b). Using ex-ante impact assessments, the adverse side ef-
fects of UPS that are invisible to the organisations implementing R & D
can be discovered during the planning process (EIARD, 2003; Helming

⁎
Corresponding author.
E-mail address: graef@zalf.de (F. Graef).

http://dx.doi.org/10.1016/j.eiar.2017.10.004
Received 11 January 2017; Received in revised form 12 September 2017; Accepted 3 October 2017
Available online 01 November 2017
0195-9255/ © 2017 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
F. Graef et al.
et al., 2011; Millstone et al., 2010). If carried out in a participatory
action research (PAR) context (Bradbury-Huang, 2010), the impact
assessment activities and interactions involving researchers and stake-
holders enable and promote recurrent UPS adaptations and various
types of social learning (Blackmore, 2007). Brydon-Miller et al. (2003),
however, warn that the feasibility of PAR is limited in a large-scale
multidisciplinary and multicultural context.
Ex-ante impact assessments have increased in importance because
they allow funding organisations to control beforehand whether their
funds are likely to be well spent (Hulme, 2000). Ex-post impact as-
sessments can be used for participatory co-learning, monitoring and
evaluating implemented UPS, and for adapting them to local require-
ments when needed (Bradbury-Huang, 2010; Graef et al., 2014; König
et al., 2013; Mayoux and Chambers, 2005). They can be done at a
specific point in time or in intervals both during and after UPS im-
plementation.
However, in implementing UPS, it is important to incorporate
gender considerations so that UPS can be adapted to the needs of men
and women in the FVC (FAO, 2011; Okonya and Kroschel, 2014; Polar
et al., 2015). As shown in various studies, R & D does not necessarily
result in greater gender equality (Arora-Jonsson, 2014; Polar et al.,
2015) and more efficient resource management (Ochola et al., 2010).
According to Arora-Jonsson (2014) each context requires “efforts to
define what gender is”. Mayoux and Chambers (2005) emphasise in
their “impact assessment agenda for pro-poor development and im-
proving practices” that both genders must participate to prioritise their
livelihood affairs. Female farmers in sub-Saharan Africa, who have la-
bour market participation shares of 60–80% in the agricultural sector
(Emerole et al., 2014; FAO, 2011; Mnimbo et al., 2015), are likely to be
more concerned about the implementation of a new R & D project than
men, particularly if it entails a potential increase in the labour required,
because women in developing countries tend to work far more hours
than men (FAO, 2009; Tsikata and Yaro, 2014). Many studies (FAO,
2009, 2011; Mayoux and Chambers, 2005; Mnimbo et al., 2015, 2017)
emphasise gender differences in rural lives, experiences, needs and
priorities and their variations depending on age, ethnicity, disability,
income levels, and marital status. They recommend applying gender-
focussed analysis frameworks that answer questions about influencing
factors such as “What is getting better? What is getting worse? Who
does/has/needs what? What does an innovation deliver to narrow the
gender gap?” (Ochola et al., 2010).
Targeting female-male differences in knowledge, perception, and
impact assessments of UPS for enhancing food security, therefore, is of
utmost importance not only at the local implementation level of non-
scientific FVC stakeholders (Grimble and Wellard, 1997) but also at the
level of scientific experts and extension staff (Bradbury-Huang, 2010,
2013; Croppenstedt et al., 2013; Graef and Uckert, 2016). Whereas
Arora-Jonsson (2014) found gender and gendered research as being
only partially transformative, Bradbury-Huang (2013) drew a picture of
more holistic post-patriarchal science in PAR that results in “partner-
ship between women and men, scholars and practitioners, university
faculty and community stakeholders”.
How can this gender perspective be included in a systematic way in
ex-ante and ex-post impact assessments of food security? This study
advances the previous participatory and methodological work of
Schindler et al. (2016a, 2016b) that, based on gender-balanced focus
group discussions, identified the locally relevant food security criteria
and indicators to be assessed and that developed a “methodological
approach used for ex-ante UPS sustainability impact assessments based
on the Framework for Participatory Impact Assessment (FoPIA)”
(Morris et al., 2011). Schindler et al. (2016c), found that the obvious
differences “between stakeholders' and researchers' knowledge can
enhance the quality of impact assessments if they are used in a com-
plementary way (Bradbury-Huang, 2010).” Schindler et al. (2016a,
2016b, 2016c) did not present a systematic gender-segregated differ-
entiation, but underlined that “assessment rounds be organized
82
Environmental Impact Assessment Review 68 (2018) 81–89
separately for male and female participants” (Schindler et al., 2016a),
because gender roles and respective feedback narratives in the rural
context differed largely. Graef and Uckert (2016), focusing exclusively
on scientists' expertise and perceptions, found gender-specific differ-
ences in ex-ante suitability assessments depending on the type of UPS
and food security criteria assessed. However, they did not reach general
conclusions about fields of male and female expertise, their social roles,
and their perceptions. In the same study context, Mnimbo et al. (2017)
found significant differences in crop and UPS preferences between
youth, women and men and recommended that a “site-specific gen-
dered analysis … in agricultural value chains should be completed prior
to introducing an intervention”.
This study adds to and combines previous sustainability impact
assessment and gender studies by providing a) a gendered analysis on
UPS implementations assessed by stakeholders for their impacts on food
security and b) by performing this analysis both ex-ante before im-
plementation and ex-post 14 months after implementation. These are
suggested as components of gender analysis frameworks (Ochola et al.,
2010). We hypothesised that differences would exist in assessments a)
between the genders, b) between the points in time of assessments, and
c) in changes in male-female perception over time. The objective of our
work was to verify the hypotheses and investigate other specific dif-
ferences, for instance, in UPS and/or food security criteria assessed. We
applied this gendered analysis with respect to social, environmental,
and economic aspects of villages in rural Tanzania.
2. Methods
2.1. Study area, food systems, and local social context
This study was carried out in four rural Tanzanian villages located
in two agro-climatically differing regions of Tanzania: the pre-
dominantly sub-humid (600–800 mm) Morogoro Region and semi-arid
(350–500 mm) Dodoma Region, together representing most of the
variability in Tanzanian farming systems (USAID, 2008). In the sub-
humid Morogoro region, which has both food-insecure and food-secure
areas, food systems are primarily based on maize, legumes, sorghum,
rice, and horticulture. Land pressure in that multi-ethnic region is high.
In the mainly food-insecure semi-arid Dodoma region, food systems
depend more on sorghum and millet, and often include livestock
(Mnenwa and Maliti, 2010; Liwenga, 2003). In this mono-ethnic region,
the land pressure is medium to high.
The main criteria for selecting the four villages included the fol-
lowing: a) similar climates; b) differing market access; c) differing
rainfed cropping systems, possibly integrating livestock; d) village sizes
of 800–1500 households, and e) high number of stunted children below
5 years of age as an indicator of food insecurity (Graef et al., 2014).
Agriculture is the overly dominant pillar of the villages' economy and
provides employment to most of its citizens, who are primarily rural
poor smallholder farmers, with a (very) few millers, traders and food
processors.
Generally, women experience more challenges in their multiple
roles, such as reproduction and food production and preparation
(Mnimbo et al., 2015). Income-generation strategies preferred by men
are much different from those preferred by women and youth, and they
are closely linked to access to and ownership of land and other re-
sources and participation in income-generating activities. Generally in
the study villages men take over all of the important decisions in the
household including what crops to produce, how much to sell, when
and where to sell (Mnimbo et al., 2017). The women consider them-
selves to be inferior to men, for instance, a women said “here in
Chamwino our men do not do most of the land tilling like we women do but
they make all the important decisions including taking the income accrued
from the crop produce, it doesn't matter if he participated during the pro-
duction or not”(female focus group participant at Ilolo village in March
2014). In the Dodoma region due to the patrilineal society like Gogo,
F. Graef et al.
Ngindo and Sukuma the position of women in decision making, and
participation in lucrative nosectionsde of the FVC is minimal. A female
key informant from Ilakala village states that “during harvesting most
men help with transporting maize if they have a means of transport, for
example, bicycle or motorbike. They will initially assist us with harvesting
but as soon as enough load for the bicycle or motorbike has been obtained,
they will completely shift to transportation and leave behind the harvesting
task to us and our kids”. In terms of household budget the income from
crop sales goes to the male. Female farmers can gain some profit by
selling local brew, as a woman from Idifu village (February 2014) states
“in order to add value to our millet we had to convert it into and sell it as
local brew. ….Normally, we divide the profit more or less on equal basis
between the husband and wife” (Mnimbo et al., 2017).
2.2. Participatory action research approach and the UPS selection
PAR is a collaborative research approach that emphasises collective
inquiry, evolving experimentation and following reflection among
those who participate as co-researchers (Bradbury-Huang, 2010). In
PAR the research activities and methodology are context-oriented and
iterative with regard to how both local stakeholders and scientists de-
velop and select research methods, generate data, and reflect in cycles
on how the efforts unfold and what are the impacts of an intervention
(Chambers 1994; Prowse 2010). If the different learning contexts are
understood by both local stakeholders and scientists, social learning
may become more valued and facilitated (Blackmore, 2007).
For the ultimate goal of enhancing rural Tanzanian food systems the
PAR approach was considered appropriate in order to include local
stakeholders knowledge, develop adapted local UPS, and create own-
ership for the UPS. In this PAR the stakeholders from the beginning
were considered as partners and co-generators of research processes
and knowledge with concerns that are valuable to their local settings
(Bradbury-Huang, 2010; Prowse 2010).
In a participatory process reported by Kaburire et al. (2015) and
Mwinuka et al. (2015), food-securing UPS were selected; both local
subsistence stakeholders and Tanzanian and German scientists with
expertise in FVC sectors were involved (Fig. 1). This iterative process
involved a) the screening of existing UPS in the case study villages, the
two focal regions and beyond and making an UPS inventory, b) an
expert-based specification and prioritisation of a few UPS, and finally c)
a stakeholder-based prioritisation of 13 UPS for implementation and
testing. Different participative methodologies were used for most of the
Fig. 1. PAR steps, stakeholder categories involved, and types of stakeholder involvements
(FGDs: focus group discussions; CSS: case study site; UPS: upgrading strategy; FVC: food
value chain), with some steps being recurrent leading to iterative changes of the following
ones.
83
Environmental Impact Assessment Review 68 (2018) 81–89
Table 1
Upgrading strategies per food value chain component, and implementation time range.
(Adapted from Schindler et al., 2016a and Trans-SEC, 2016).
Natural resource management/crop production
Rainwater harvesting and micro-dosing fertilization (RWH/MF): In situ RWH using
tied ridges in the sub-humid region and infiltration pits in the semi-arid region
(Mahoo et al., 2012); micro-dose rates of 5–6 kg P/ha (2–5 g/hill as NPK) placed
4–8 cm from and lateral to the seeds, with higher rates in more humid climates
(Bagayoko et al., 2011); implementation time range of 3–5 months; started in
December 2014.
Post-harvesting processing and biomass/energy supply
Improved cooking stoves (ICS): Small-scale stoves from loam, reducing energy
consumed for household use with two holes and smoke exhaustion via a chimney
at 3–5 US$/stove, locally constructed by trainers training other stakeholders
(Kshirsagar and Kalamkar, 2014); implementation time range of 6–8 months;
started in February 2015.
Consumption
Household nutrition education and kitchen garden training (HNE/KG): Increase
awareness of nutrient-rich indigenous foods and making better use of crops (Roy
et al., 2005); cultivating indigenous fruits and vegetables for dietary
diversification (Galhena et al., 2013); implementation time range of 3–6 months;
started in December 2014.
analytical steps (Fig. 1). Participation intensity levels included in-
formation, consultation, cooperation, collaboration, and empowerment
(Stauffacher et al., 2008). The successive co-generation of stakeholders
and scientists' knowledge iteratively and in a recurrent approach
shaped most of those methodologies (Bradbury-Huang, 2010).
Among the 13 prioritised UPS, the authors selected three low-input
UPS for presenting these ex-ante and ex-post impact assessments
(Table 1): 1) rainwater harvesting and micro-dosing fertilization
(RWH/MF), 2) improved cooking stoves (ICS), and 3) household nu-
trition education and kitchen garden (HNE/KG). All of these UPS re-
quire only a small financial input and were implemented early (within a
range of 3–8 months) after their final participative prioritisation and
announcement in September 2014 (Schindler et al., 2016b) (Table 1).
2.3. The modified Framework for Participatory Impact Assessment
The methodological framework used for the UPS impact assess-
ments was based on the FoPIA (Framework for Participatory Impact
Assessment) approach (Morris et al., 2011), which was originally de-
veloped to link computer-based sustainability impact assessments with
qualitative participatory approaches (Helming et al., 2011). The FoPIA
was modified by König et al. (2013) to be used in different contexts on
the regional policy scale of poor countries. The FoPIA was further de-
veloped by Schindler et al. (2016a, 2016b) and applied at the rural
village community level to assess food-securing UPS prior to their im-
plementation (ex-ante, 2014). For that purpose, a balanced community-
based set of nine food security criteria was co-developed with local
stakeholders (Table 2). These criteria were balanced along the sus-
tainability dimensions (economic, environmental, social) while re-
presenting the four food security dimensions (food availability, access,
utilisation, and stability). We considered these food security criteria
suitable for assessing the impact of the food-securing UPS. For this
study, the methodological approach of Schindler et al. (2016a, 2016b)
was enlarged to include ex-post impact assessments by local stake-
holders across the food value chains to enable impact assessment both
before (T0) and after (T1) the UPS implementation. Furthermore, we
included a gendered framework and analysis that enables us to target
the main UPS challenges as experienced separately by female and male
stakeholders, providing an opportunity to take corrective action to
adjust possible shortcomings (Silvestrini, 2011). Finally, we evaluate
the food security impacts experienced thus far. The methodological
approach used involves various successive mixed-gendered focus group
workshops, is comprehensive and considers the local socio-cultural
setting (Mayoux and Chambers, 2005; Schindler et al., 2016a, 2016b).
F. Graef et al. Environmental Impact Assessment Review 68 (2018) 81–89

Table 2
Food security criteria and explanation.
(Adapted from Schindler et al., 2016b).

Criterion Sustainability dimension Definition and regional relevance

Food diversity Social Sufficient number of meals (= 3) per day offering a diversified and balanced diet
Social relations Social Community support during family need (i.e., drought, family issues such as illness and death) and share of the workload (i.e.,
field ploughing); family support and understanding of decision making about household resources
Working conditions Social Access to appropriate technology/equipment and agricultural practices, reducing working hours and workload
Yield Economic Amount of food produced and available for family consumption and for sale
Income Economic Family financial resources earned from agricultural production and off-farm activities
Market participation Economic Selling and buying agricultural products and other needs; knowledge of market prices for improved negotiation power of
farmers towards buyers
Soil fertility Environmental Quality of the soil for agricultural production
Water availability Environmental Soil water availability for agricultural production
Agro-diversity Environmental Cultivation of different crops for family consumption and for selling; risk management in the case of one crop failure

Villagers were informed about this type of PAR and after providing
their consent to participate as co-researchers, a set of stakeholders re-
presenting a socio-economic cross-section of the village was selected for
focus group discussions (FGD) (Schindler et al., 2016b).
We used a bipolar Likert assessment scale (Harpe, 2015) ranging
from − 3 to + 3 (− 3 very high negative impact, −2 medium negative
impact, − 1 minor negative impact, 0 no impact, + 1 minor positive
impact, +2 medium positive impact, + 3 high positive impact). The
stakeholders were asked in separate FGD ex-ante directly after UPS
selection but prior to implementation to assess how the UPS were ex-
pected to affect food security and its related criteria (Table 2). The FDG
participants discussed the UPS impacts together, produced secret (un-
biased) scores that were averaged per criterion and presented the re-
sults to the group. Thereafter, a few individuals were interviewed by a
same-gender scientist to gain more specific, but possibly hidden, in-
formation. Fourteen months after the UPS were implemented, the sta-
keholder groups were asked again to assess how the UPS affected the
food security criteria.
included different levels of comparison. We investigated all levels for
gender-specific differences and/or similarities, scoring ranges (Min/
Max values), and standard deviations. The gender comparisons by re-
gion and/or village produced less significant results, most likely be-
cause of the low number of assessing stakeholders per UPS and village,
9–14. Therefore, for male versus female and for T0 versus T1 compar-
isons, the assessment data for male and female stakeholders were se-
parately pooled across all villages and regions (Table 3), although
Schindler et al. (2016a, 2016b) and Hernandez (2016) indicated that
this could mask region- or village-specific differences and even negative
assessments.
The data were analysed with IBM SPSS Statistics 22 using de-
scriptive statistics and the Mann-Whitney U test for non-parametric
rating scale data to test for significant differences between the genders
and the two points in time (Harpe, 2015). For a better interpretation
and differentiation of the scoring results, we calculated and presented
the arithmetic averages and standard deviations instead of using the
median values, which are used in the Mann-Whitney U test.

2.4. Gendered analytical framework and data analysis 3. Results

Many of the development inequalities emerge from gender differ-
ences (White et al., 2016). Also, there is limited literature linking FVC
and gender (Bolwig et al., 2010). Since gender is one of the most im-
portant aspects in sustainable development, involving participation of
men and women in the whole process of FVC development is key (Thow
et al., 2017). Particular attention is required on the analysis of the
horizontal relationships between actors in the FVC (Norell et al., 2016),
so that UPS do not overlook gender relations and negatively impact
women and youth (Tsikata and Yaro, 2014). In another study we found
differences between Tanzanian men versus women and youth in the
UPS preferences across the FVC, indicating their distinctions in FVC
participation (Mnimbo et al., 2017). In rural Tanzania youth and
women were more involved with crop harvesting, transportation and
primary processing, while men were more engaged with handling farm
inputs and crop marketing.
Our research included all elements of the Harvard analytical fra-
mework also known as the Gender Roles Framework (Cole et al., 2015).
This framework analyses the activity profiles (who does what?); access
and control profiles (who controls what?); influencing factors (who is
influenced by what?); and the project cycle (at which point are women
and/or men involved with the project)? The latter was found particu-
larly important because the UPS participation changed with time
among the gender groups.
The PAR activities (team formation, planning, execution, experience
sharing) towards UPS implementation and impact assessments in many
aspects included a gendered perspective with both scientists and sta-
keholders acting either in mixed (FDGs) or unisex (interviews) teams
(Fig. 2). The impact assessment scores were analysed in a gender-seg-
regated way following a hierarchical analytical framework that
The UPS impact assessments of the local food security criteria only
differed significantly in a few cases (p < 0.05) between the genders of
stakeholders (Table 3). Although many of the assessments differed to a
large degree (> 0.5 score difference), in most cases this was statisti-
cally insignificant, probably because of either the high disunity among
assessments, mirrored by high SDs, or the limited number of assessors.
Cases with high disunity (SD > 1.0) were found among both genders
and among only men or only women per criterion assessed. The genders
did not differ overall in minimum and maximum scorings. Overall, fe-
male stakeholders gave slightly lower ratings across all three sustain-
ability dimensions. The analysis revealed that men more often revised
high ratings from T0 to T1. This could be explained by the fact that men
are more self-confident when discussing knowledge-based issues and
expectations (T0, ex-ante), whereas women become more self-confident
when referring to their own experiences in T1 (ex-post). This was
particularly the case in situations in which women obviously have the
major burden, such as cooking, firewood collection and kitchen gar-
dens.
Ex-ante (T0) versus ex-post (T1) assessment comparisons exhibited
overall high (p < 0.05) to highly (p < 0.001) significant differences
across all three UPS among both male and female stakeholders
(Table 3). The standard deviations were higher in T0 than in T1, in-
dicating an equalisation of prior expected high or low impacts with
time and experience. In most cases, the assessments distinctly increased
from ex-ante to ex-post for ICS, whereas for RWH/MF and HNE/KG, the
opposite was observed. For RWH/MF, the significant T0/T1 differences
matched well between the genders except for working conditions, which
produced a highly significant T0/T1 decrease exclusively among fe-
males. Interestingly, for ICS and HNE/KG impact assessments, the

84
F. Graef et al. Environmental Impact Assessment Review 68 (2018) 81–89

Fig. 2. Gender-focussed impact assessments framework and

analysis: PAR activities related to UPS implementation and
impact assessments are gendered among both scientists and
stakeholders and interact through cyclic reflection. Impact
assessment scores are investigated at different levels of
comparison starting from ex-ante (T0) versus ex-post as-
sessments (T1) down to the food security criteria; region-
and village-specific differences among male and female
stakeholders were not analysed in this study.

significant T0/T1 differences often did not coincide between the gen- available soil water (both increased), whereas for female stakeholders,
ders. For instance, for ICS, only male stakeholders' T0/T1 differences there was a significant T0/T1 increase for yield, social relations and agro-
were significant for market participation (decrease), soil fertility and diversity. For HNE/KG, only male stakeholders' T0/T1 differences were

Table 3
Differences in food security assessment criteria between male and female stakeholders at T0 (ex-ante right after UPS implementation decision) and T1 (ex-post assessment 13 months after
UPS decision making).

Ex-ante (T0) Ex-post (T1)

Mean SD (N) Mean SD (N) Mean SD (N) Mean SD (N) Mean SD (N) Mean SD (N)

Economic criteria Production Income Market participation Production Income Market participation

Rainwater harvesting, M 2.33* 1.02 24 2.29** 1.07 24 2.00 1.01 24 1.48* 1.48 28 1.33** 1.30 28 1.41 1.15 28
fertilizer micro- F 2.52* 0.81 21 2.19** 1.11 21 1.76 0.98 21 1.71* 1.31 21 0.95** 1.47 21 1.43 1.50 21
dosing
Improved cooking M 1.14 1.41 14 1.00 1.30 14 2.36* 0.84 14 1.53 1.41 18 1.40 1.55 18 1.00* 1.69 18
stove F 0.67* 1.32 9 1.33 1.58 9 1.67 1.41 9 2.00* 1.06 27 1.85 1.29 27 1.46 1.24 27
Nutrition education, M 2.95*** 0.23 19 2.11 1.59 19 2.68* 0.58 19 1.67*** 1.19 19 1.61 1.20 19 1.56* 1.58 19
kitchen gardens F 2.45 1.01 22 2.86* 0.35 22 2.73** 0.77 22 2.29 0.95 26 2.13* 1.26 26 1.79** 1.44 26

Social criteria Food diversity Social relations Working conditions Food diversity Social relations Working conditions
Rainwater harvesting, M 1.90* 1.03 24 2.08 1.36 24 1.98 0.97 24 1.26* 0.90 28 2.07 1.21 28 1.37 1.52 28
fertilizer micro- F 1.90* 1.01 21 1.90 1.18 21 2.21** 0.90 21 1.05* 1.28 21 1.76 1.14 21 0.95** 1.60 21
dosing
Improved cooking M 0.50 0.94 14 1.14 1.35 14 0.21*** 0.80 14 1.13+ 1.46 18 2.13 1.19 18 2.40*** 1.06 18
stove F 1.22 1.48 9 1.33* 1.58 9 0.33*** 1.00 9 2.27+ 1.04 27 2.65* 0.85 27 2.65*** 0.63 27
Nutrition education, M 2.26+ 1.05 19 2.68 0.75 19 2.37* 0.90 19 1.94+ 1.00 19 2.39 0.85 19 1.11+⁎ 1.88 19
kitchen gardens F 2.82+ 0.66 22 2.32 1.09 22 2.36 1.29 22 2.54+ 0.78 26 2.25 1.29 26 2.25+ 1.11 26

Environmental criteria Soil fertility Available soil water Agro-diversity Soil fertility Available soil water Agro-diversity
Rainwater harvesting, M 2.65** 0.58 24 2.04 1.22 24 2.60 0.88 24 2.00** 0.83 28 2.33 0.73 28 2.44 0.89 28
fertilizer micro- F 2.71** 0.54 21 2.12 1.11 21 2.62 0.71 21 1.95** 1.12 21 2.00 1.00 21 2.33 0.97 21
dosing
Improved cooking M 0.00*** 0.00 14 0.00** 0.00 14 0.43 0.94 14 1.13*** 1.19 18 0.73** 1.44 18 1.07 1.71 18
stove F 0.00 0.00 9 0.00 0.00 9 0.00*** 0.00 9 0.77 1.34 27 0.73 1.19 27 1.62*** 1.27 27
Nutrition education, M 0.74 2.16 19 2.11++** 1.33 19 3.00*** 0.00 19 1.17 1.62 19 0.61** 2.17 19 2.17*** 1.04 19
kitchen gardens F 0.68* 1.29 22 0.68++ 1.29 22 2.91*** 0.43 22 1.54* 1.67 26 1.29 1.76 26 2.13*** 1.08 26

Differences between male and female stakeholders at significance levels of + p < 0.05, ++ p < 0.01; differences between T0 and T1 at significance levels of *p < 0.05, **p < 0.01,
***p < 0.001 with T0 = ex-ante assessment right after UPS implementation decision and T1 = ex-post assessment 14 months after UPS decision-making.

85
F. Graef et al.
significant for yield, working conditions and available soil water, whereas
for female stakeholders, there was a significant T0/T1 soil fertility as-
sessment increase. The most significant T0/T1 assessment increases
overall were found for both genders for ICS impacts on working condi-
tions and social relations and for males only on yield and agro-diversity.
The most significant T0/T1 assessment decline overall was found for
both genders for the RWH/MF impact on yield and income and for fe-
males only on working conditions. These gender-specific T0/T1 differ-
ences indicate that both genders had different UPS-specific and criteria-
specific expectations ex-ante and experiences ex-post for the UPS im-
plementation. Although the trend of increase and decline from T0 to T1
coincided well between the genders in most cases, the differences in the
genders' assessments were considerable. This shows that both the per-
ception at a specific point in time and the perception change over time
of male and female stakeholders is a critical factor in assessing food-
securing UPS.
A comparison of the different UPS by averaging the scores across all
food security criteria exhibited ex-ante scores of overall moderate-to-
high positive food security impacts for RWH/MF and HNE/KG and only
low positive impacts for ICS, with males scoring distinctly lower. The
averaged ex-post assessments for all three UPS produced similar mod-
erately positive scores of 1.7 to 2.0, with RWH/MF and HNE/KG de-
creasing and ICS increasing compared to T0.
Comparing the FS criteria in economic, social and environmental
sustainability dimensions (Tables 2 and 3), it was found that gender
differences are more pronounced among the social criteria, whereas
T0/T1 differences exist in all three dimensions. Overall, economic and
social impacts scored higher than environmental impacts.
The overall highest ex-ante scores per criteria (> 2.5) were attrib-
uted to the HNE/KG impact on yield, income, market participation, food
diversity, social relations and agro-diversity and to the RWH/MF impact
on soil fertility and agro-diversity. The highest ex-post scores (> 2.5)
were given only to ICS on social relations and working conditions. Zero
scores were given ex-ante to ICS impacts on soil fertility, available soil
water, and agro-diversity, all of which slightly increased during ex-post
assessments.
4. Discussion
4.1. Gendered stakeholder assessments (ex-ante versus ex-post)
The aim of our gender-segregated participatory impact assessments
of social, economic and environmental sustainability was to derive in-
dications of potential successes and challenges and to uncover un-
expected themes before and during UPS implementation (Grimble and
Wellard, 1997; Hernandez, 2016; Riisgaard et al., 2010) as perceived by
the female and male stakeholders involved (Croppenstedt et al., 2013;
Mnimbo et al., 2015; Mnimbo et al., 2017). This could serve to fine-tune
and adapt the UPS during implementation with the intent to achieve
greater sustainability and gender equity (Polar et al., 2015). As shown
in Fig. 2, we systematically studied our gender-specific findings at
different scales and content levels.
The food security impact assessment results were regionally pooled
and gender segregated. It was found that simultaneous consideration of
the nine food security criteria among the three sustainability dimen-
sions was pivotal to achieve a healthy equilibrium (Helming et al.,
2011) and enable sustainable local food security for both genders,
particularly in the small-scale farming setting (Mnimbo et al., 2017;
Zimmerer et al., 2015). Our assessments, particularly our ex-ante as-
sessments, showed that the positive impacts were overly dominant,
with only a few zero impact cases indicated (Table 3). Some reasons for
this finding could be that a) the UPS had been already ex-ante reviewed
and evaluated by scientists, supported by the literature, b) the UPS were
then chosen by the stakeholders based on that information, and c) the
scientists attending the assessment exercise may have prevented the
stakeholders from providing negative UPS aspects (Jakobsen, 2012) for
86
Environmental Impact Assessment Review 68 (2018) 81–89
fear of losing the project's support, a situation that is often observed in
NGO projects offering benefits. The decrease in assessment range from
T0 to T1 is clear evidence of how the pool of opinions homogenised,
and it shows the trade-offs between negative and positive UPS impacts
(Hernandez, 2016).
Many characteristic and sometimes significant scoring differences
were found between the genders of stakeholders and the assessment
points in time (ex-ante versus ex-post). For instance, the positive ex-
ante assessments of RWH/MF impacts by both genders were attributed
to increased agricultural production yield, expected increased income
and, therefore, better options for diversifying the diet. However, the
decline in the ex-post assessments and the stakeholders' narratives in-
dicated some disillusionment caused by rainfall uncertainty and the
increased workload of tied ridge construction (Table 3), because ac-
cording to the FAO (2011), women have higher labour shares in most of
Sub-Saharan Africa. Croppenstedt et al. (2013) and Mnimbo et al.
(2017) found that the higher workload affected women more than men.
According to Schindler et al. (2016a), fertilizer micro-dosing was a
subject of controversy among both genders, indicating “the need for
farmers' training and knowledge sharing in regard to fertilizer place-
ment and timing and the advantages and risks of fertilizer use” (Chianu
et al., 2012).
Retrospectively, female participants of the FoPIA workshops dis-
cussed more of their experience in ex-post assessments (T1), indicating
higher positive impacts on working conditions for ICS and HNE/KG
compared to male participants, showing that women tend to be more
self-confident when referring to their own experiences (Arora-Jonsson,
2014; FAO, 2011; Mnimbo et al., 2017).
Therefore, gender, age and income-specific roles, power relation-
ships, and priorities play an important role because field ownership,
crop type, fieldwork hours invested, and education are dis-
proportionally distributed among male and female farmers (Mnimbo
et al., 2017; FAO, 2009; Ochola et al., 2010; Polar et al., 2015; Tsikata
and Yaro, 2014). Arora-Jonsson (2014) noted that women often ex-
perience “intersecting structural constraints” and that “richer analyses
of women's and men's access, ownership, use, rights or action over
environmental resources is enabled when focusing not only on gender
but also in relation to other cross cutting categories that intersect with
social positions as women or men and their class, education, and eth-
nicity…”.
Because stakeholders are involved in many different ways in dif-
ferent FVC sectors, the introduction of UPS is expected to affect their
participation in different ways that may also change with time and
implementation stage. Kowalski et al. (2015) found four opportunities
of influence of the UPS: inclusion, repositioning, expulsion, and non-
participation. Considering the high local and temporal variability in
FVC sector participation we recommend that gendered analysis on
potential UPS across FVC should be site-specific and done before in-
tervention.
4.2. Key gendered findings of the PAR approach (scientists and
stakeholders)
Understanding the aforementioned intersecting structural con-
straints of the genders becomes even more complex in a multi-cultural
and multi-disciplinary PAR setting, as we find in this project (Graef
et al., 2014; Popescu et al. 2014). As explained above, the PAR ap-
proach to UPS implementation and impact assessment was largely
gender-focussed, as illustrated in Fig. 2, although it was far from being
conceptualised as non-patriarchal (Arora-Jonsson, 2014; Bradbury-
Huang, 2013). Although there were fewer female scientists in most PAR
activities (Graef and Uckert, 2016), they were present in both impact
assessments both to provide a trusting environment that empowered
local women to share their knowledge (Aziz et al., 2011) and to gain
better gendered insights into local food security challenges (Ochola
et al., 2010). A cross-cutting issue during implementation was the
F. Graef et al.
indication of potentially increasing tensions caused by jealousy among
non-implementers and implementers of UPS, among the UPS groups,
and between genders, all of which affected stakeholders' social rela-
tions, one of the food security criteria (Table 3). Scientists and devel-
opers involved in the process, therefore, must not only consider this
issue during the UPS implementation (FAO, 2011; Mayoux and
Chambers, 2005; Polar et al., 2015) and enable more open access by all
village stakeholders to knowledge, resources and technologies but also
create better connections between male and female stakeholders and
scientists to achieve better knowledge of the complex local food system
and local social relations (Arora-Jonsson, 2014; Reed, 2008; Millstone
et al., 2010).
Although the food security impacts of UPS were perceived differ-
ently between the female and male stakeholders involved
(Croppenstedt et al., 2013), male and female scientist counterparts in
another study also differed in their ex-ante suitability assessments
(Schindler et al., 2016c; Graef and Uckert, 2016). Reflective exploration
of food security impacts, their assessment, and perception differences
between male and female farmers and scientists as core PAR features
were found to be very valuable in improving mutual social learning for
implementing more sustainable solutions (Bradbury-Huang, 2010;
Blackmore, 2007). Schindler et al. (2016c), found that “farmers and
scientists have considerably different views on the positive and negative
impacts of UPS implementation, with scientists mostly focusing on di-
rect causal impact chains and the farmers tending to consider indirect
linkages that affect their complex livelihoods” (Bond and Morrison-
Saunders, 2011; Tittonell et al., 2007). Multiple gender-, discipline- and
culture-specific perspectives raised questions and awareness about the
nature of knowledge among involved participants, both scientists and
stakeholders, and boosted co-learning (Blackmore, 2007). Hence, the
PAR approach was successful in many ways and benefited the local
communities by producing 13 locally adapted food securing UPS which
up to now enhanced the livelihoods of many thousands of households.
Brydon-Miller et al. (2003), however, warned about PAR being less
successful in a larger-scale project context. In this study, for instance,
which included some highly complex PAR activities among scientists
we experienced less successful North-South collaboration, co-learning,
and transforming knowledge into actionability in the context of UPS
implementation (Bradbury-Huang, 2010; Popescu et al., 2014).
4.3. Impact assessment methodology and analytical challenges
Consideration and discussion of the views of all participants during
the FGDs was found to be fundamental (Bond et al., 2012; Jakobsen,
2012), as was collecting the gendered arguments and narratives un-
derlying the assessment scores (Schindler et al., 2016a). Therefore, both
women and men could reflect on their individual perceptions of the
negative and positive impacts of UPS on their food security situation.
This also supported both stakeholders and participating scientists “be-
coming aware of their assumptions regarding learning, knowing and
how [they] develop knowledge, and find out more about what [they]
need to know”, as Blackmore (2007) termed it.
Various trade-offs during UPS implementation were indicated
throughout all FGDs (Morris et al., 2011) and unavoidably changed
based on the point in time of the assessment (Table 3). Considering and
discussing them was key to informing the stakeholders so that they
could make decisions and compromises involving the allocation of their
available resources (Mnimbo et al., 2017; Tittonell et al., 2007). The
discovery, discussion and timely change in trade-off perceptions can
therefore be considered a co-learning experience (Giller et al., 2011).
During the implementation processes from T0 to T1, bottlenecks were
identified by both male and female stakeholders and scientists, leading
them to reflect on their knowledge gaps and co-learn what they needed
to know (Blackmore, 2007). Measures were taken to adapt the UPS to
the local environment and requirements to avoid failure (König et al.,
2013; Mayoux and Chambers, 2005), thus participatively and
87
Environmental Impact Assessment Review 68 (2018) 81–89
successively co-creating the final versions of the UPS (Bond et al., 2012;
Trans-SEC, 2016).
Because women tended to speak less during the gender-mixed FGDs
and impact assessment rounds, we recommend that future impact as-
sessments be conducted separately for men and women (Aziz et al.,
2011; Jakobsen, 2012; Kristjanson et al., 2002; Mnimbo et al., 2017).
Furthermore, a successful assessment moderation process requires both
male and female empathetic gender-sensitised moderators who are
acquainted with both local culture and language (Jakobsen, 2012;
Reed, 2008), with scientists preferably staying in the background
(Schindler et al., 2016b). In this study, the moderators made the sta-
keholders feel that they were listened to and taken seriously in their
gendered role as men or women, encouraging lively knowledge ex-
change and learning among stakeholders and scientists. They perceived
gender-sensitive topics that preferably were discussed in same-gender
FGDs or interviews (Mnimbo et al., 2017) and that were later shared
with the entire group (Schindler et al., 2016b). Gendered participatory
ex-post monitoring and evaluation while revisiting locally relevant food
security criteria should be housed as long-running programmes
throughout the duration of the UPS implementation to capture stake-
holders' perception, gender role, and UPS performance changes at later
stages, continue the co-learning process and develop the capacity to
engage with gender issues at other levels (Aziz et al., 2011).
5. Conclusion
The study presented here applied a gendered participatory impact
assessment (FoPIA) involving scientists and poor farmers in rural
Tanzania immediately before implementation and 14 months after
implementation of food-securing upgrading strategies. Both male and
female farmers assessed the positive and negative impacts on food se-
curity criteria and indicated various trade-offs between the UPS and
their livelihoods. Assessments differed largely between the genders for
a) different food security criteria, b) different sustainability dimensions
(economic, social, and environmental), c) different points in time (T0,
T1) of the assessments, d) the different implemented UPS, and e) the
members within the groups of female and/or male stakeholders.
Additionally, perceptions differed between scientists and stakeholders.
Gender-wise differentiated participatory impact assessments, such as
those presented here, enabled us to select and/or adapt the intended
development interventions to be more focussed on gender roles, gender
power relationships, and gender equitability in a specific location and
to minimise potential adverse socio-cultural side effects that may widen
the gender gap. This study encourages the planning, execution and
analysis of UPS impact assessments to be participatory, gender focussed
and gender segregated, particularly with respect to food security and
their implementation in food value chains. Because impact assessments
vary between region and locale, their results cannot be simply trans-
ferred but instead require a site-specific application, even if the distance
between locations is small. Participative stakeholder involvement and
gendered co-learning for scientists and stakeholders are essential for
sustainability impact assessments, providing better mutual and deeper
comprehension of complex rural food systems.
Acknowledgements
This publication is a product of the project Trans-SEC (www.trans-
sec.org) under grant No. 031A249A. The German Federal Ministry of
Education and Research (BMBF) funded the project, and the German
Federal Ministry for Economic Cooperation and Development (BMZ) co-
financed this work. The views expressed are purely those of the authors
and should not under any circumstances be regarded as stating an of-
ficial position of the BMBF and BMZ.
F. Graef et al.
References
Arora-Jonsson, S., 2014. Forty years of gender research and environmental policy: where
do we stand? Women's Stud. Int. Forum 47, 295–308.
Aziz, A., Shams, M., Khan, K.S., 2011. Participatory action research as the approach for
women's empowerment. Action Res. 9 (3), 303–323.
Bagayoko, M., Maman, N., Palé, S., Sirifi, S., Taonda, S., Traore, S., et al., 2011. Microdose
and N and P fertilizer application rates for pearl millet in West Africa. Afr. J. Agric.
Res. 6 (5), 1141.
Blackmore, C., 2007. What kinds of knowledge, knowing and learning are required for
addressing resource dilemmas?: a theoretical overview. Environ. Sci. Pol. 10 (6),
512–525.
Bolwig, S., Ponte, S., Du Toit, A., Riisgard, L., Halberg, N., 2010. Integrating poverty and
environmental concerns into value-chain analysis: a conceptual framework.
Developmental Policy Review 28 (2), 173–194.
Bond, A., Morrison-Saunders, A., 2011. Re-evaluating sustainability assessment: aligning
the vision and the practice. Environ. Impact Assess. Rev. 31 (1), 1–7.
Bond, A., Morrison-Saunders, A., Pope, J., 2012. Sustainability assessment: the state of
the art. Impact Assessment and Project Appraisal 30 (1), 53–62.
Bradbury-Huang, H., 2010. What is good action research? Why the resurgent interest?
Action Res. 8 (1), 93–109.
Bradbury-Huang, H., 2013. Toward a post-patriarchal science? Action Res. 11 (3),
218–219.
Brydon-Miller, M., Greenwood, D., Maguire, P., 2003. Why action research? Action Res. 1
(1), 9–28.
CGIAR, 2012. CGIAR research program on climate change, Agriculture and Food Security
(CCAFS). http://ccafs.cgiar.org/commission/reports#final.
Chianu, J., Chianu, J., Mairura, F., 2012. Mineral fertilizers in the farming systems of sub-
Saharan Africa. A review. Agron. Sustain. Dev. 32 (2), 545–566.
Chambers, R., 1994. Participatory Rural Appraisal (PRA): Analysis of experience. World
Development 22, 1253–1268.
Cole, S.M., Puskur, R., Rajaratnam, S., Zulu, F., 2015. Exploring the intricate relationship
between poverty, gender inequality and rural masculinity: A case study from an
aquatic agricultural system in Zambia. Culture, Society and Masculinities 7 (2), 154.
Croppenstedt, A., Goldstein, M., Rosas, N., 2013. Gender and agriculture: inefficiencies,
segregation, and low productivity traps. World Bank Res. Obs. 28 (1), 79–109.
EIARD, 2003. Impact assessment and evaluation in agricultural research for development.
Agric. Syst. 78 (2), 329–336. http://dx.doi.org/10.1016/S0308-521X(03)00132-X.
Emerole, C., Ihedioha, O., Anyiro, C., Nwachukwu, A., Osondu, K., Ibezim, G., 2014.
Comparative gender dimensions in food unsecured farm households in Orsu local
government area of Imo state, Nigeria. American Journal of Experimental Agriculture
4 (4), 391–404.
FAO, 2009. Facts and Figures on Gender and Food Security - Division of Labour,
Factsheet. Rome, FAO.
FAO, 2011. Women and agriculture. In: Closing the Gender Gap for Development. The
State of Food and Agriculture. FAO, Rome.
Galhena, D., Freed, R., Maredia, K., 2013. Home gardens: a promising approach to en-
hance household food security and wellbeing. Agriculture & Food Security 2 (1),
1–13. http://dx.doi.org/10.1186/2048-7010-2-8.
Giller, K.E., Tittonell, P., Rufino, M.C., van Wijk, M.T., Zingore, S., Mapfumo, P., et al.,
2011. Communicating complexity: integrated assessment of trade-offs concerning soil
fertility management within African farming systems to support innovation and de-
velopment. Agric. Syst. 104 (2), 191–203.
Gomez, M.I., Barrett, C.B., Buck, L.E., De Groote, H., Ferris, S., Gao, H.O., McCullough, E.,
Miller, D.D., Outhred, H., Pell, A.N., Reardon, T., Retnanestri, M., Ruben, R., Struebi,
P., Swinnen, J., Touesnad, M.A., Weinberger, K., Keatinge, J.D.H., Milstein, M.B.,
Yang, R.Y., 2011. Research principles for developing country food value chains.
Science 332, 1154–1155.
Graef, F., Uckert, G., 2016. Male and female scientists differ in their suitability assess-
ments of food-securing upgrading strategies. Int. J. Agric. Resour. Gov. Ecol. 12 (3),
305–326.
Graef, F., Sieber, S., Mutabazi, K., Asch, F., Biesalski, H.K., Bitegeko, J., et al., 2014.
Framework for participatory food security research in rural food value chains. Global
Food Security 3 (1), 8–15.
Grimble, R.J., Wellard, K., 1997. Stakeholder methodologies in natural resource man-
agement: a review of principles, contexts, experiences and opportunities. ODA so-
cioeconomics workshop. Agric. Syst. 55 (2), 173–193.
Harpe, S.E., 2015. How to analyze Likert and other rating scale data. Currents in
Pharmacy Teaching and Learning 7 (6), 836–850.
Helming, K., Diehl, K., Kuhlman, T., Jansson, T., Verburg, P.H., Bakker, M., et al., 2011.
Ex ante impact assessment of policies affecting Land use, Part B: application of the
analytical Framework. Ecol. Soc. 16 (1).
Hernandez, L.E.A., 2016. Impact assessement of upgrading strategies for food security:
case study Tanzania. (MSc thesis at Humboldt University of Berlin. 169 p).
Hulme, D., 2000. Impact assessment methodologies for microfinance: theory, experience
and better practice. World Dev. 28 (1), 79–98. http://dx.doi.org/10.1016/S0305-
750X(99)00119-9.
IAASTD, 2009. Agriculture at a Crossroads: Global Report.
Jakobsen, H., 2012. Focus groups and methodological rigour outside the minority world:
making the method work to its strengths in Tanzania. Qual. Res. 12 (2), 111–130.
Kaburire, L., Graef, F., Mutabazi, K., Makoko, B., Swai, 2015. Multistakeholder platforms
for sustainable upgrading agri-food value chains in rural Tanzania: creating a space
for empowerment and effective engagement. In: Plant2030 Status Seminar. German
Federal Ministry of Education and Research (BMBF) (Berlin 4th–6th March 2015).
König, H.J., Uthes, S., Schuler, J., Zhen, L., Purushothaman, S., Suarma, U., et al., 2013.
88
Environmental Impact Assessment Review 68 (2018) 81–89
Regional impact assessment of land use scenarios in developing countries using the
FoPIA approach: findings from five case studies. J. Environ. Manag. 127 (Suppl. (0)),
56–64.
Kowalski, P., et al., 2015. Participation of developing countries in global value chains:
implications for trade and trade-related policies. In: OECD Trade Policy Papers, No.
179. OECD Publishing, Paris.
Kristjanson, P., Place, F., Franzel, S., Thornton, P.K., 2002. Assessing research impact on
poverty: the importance of farmers' perspectives. Agric. Syst. 72 (1), 73–92.
Kshirsagar, M.P., Kalamkar, V.R., 2014. A comprehensive review on biomass cookstoves
and a systematic approach for modern cookstove design. Renew. Sust. Energ. Rev. 30
(0), 580–603.
Liwenga, E., 2003. Food Insecurity and Coping Strategies in Semiarid Areas: The Case of
Mvumi in Central Tanzania. Stockholm University, Stockholm.
Mahoo, H., Kahimba, F., Mutabazi, K., Tumbo, S., Rwehumbiza, F., Reuben, P., et al.,
2012. Adoption and up scaling of water harvesting technologies in Tanzania. In:
Christley, W., Goeing, J. (Eds.), Water Harvesting Technologies in SSA: State of the
Art. Earth Scan publishers.
Mayoux, L., Chambers, R., 2005. Reversing the paradigm: quantification, participatory
methods and pro-poor impact assessment. J. Int. Dev. 17 (2), 271–298.
Millennium Villages, 2013. http://www.millenniumvillages.org/the-villages.
Millstone, E., Van Zwanenberg, P., Marshall, F., 2010. Monitoring and evaluating agri-
cultural science and technology projects: theories, practices and problems. IDS Bull.
41 (6), 75–87. http://dx.doi.org/10.1111/j.1759-5436.2010.00185.x.
Mnenwa, R., Maliti, E., 2010. A comparative analysis of poverty incidence in farming
systems of Tanzania. In: Special Paper 10/4. Dar es Salaam: Research on Poverty
Alleviation (REPOA).
Mnimbo, T.S., Mbwambo, J., Kahimba, F.C., Tumbo, S.D., 2015. A gendered analysis of
perception and vulnerability to climate change among smallholder farmers: the case
of Same District, Tanzania. Climate and Development. http://dx.doi.org/10.1080/
17565529.2015.1005038.
Mnimbo, T.S., Lyimo-Macha, J., Urassa, J.K., Mahoo, H.F., Tumbo, S.D., Graef, F., 2017.
Influence of gender on roles, choices of crop types and value chain upgrading stra-
tegies in semi-Arid and sub-humid Tanzania. Food Security. http://dx.doi.org/10.
1007/s12571-017-0682-2.
Morris, J., Tassone, V., Groot, R.d., Camilleri, M., Moncada, S., 2011. A framework for
participatory impact assessment: involving stakeholders in european policy making, a
case study of land use change in Malta. Ecol. Soc. 16 (1).
Morrison-Saunders, A., Pope, J., Bond, A., Retief, F., 2014. Towards sustainability as-
sessment follow-up. Environ. Impact Assess. Rev. 45, 38–45.
Mwinuka, L., Schneider, I., Maeda, C., Mutabazi, K.D., Makindara, J., Graef, F., Sieber, S.,
Swai, E., Mbwana, H., Swamila, M., 2015. Comparing stakeholder views for mutual
acceptable food value chain upgrading strategies in Tanzania. Afr. J. Agric. Res. 10
(12), 1376–1385.
Norell, D., Janoch, E., Kaganzi, E., Tolat, M., Lynn, M.L., Riley, E.C., 2016. Value chain
development with the extremely poor: evidence and lessons from CARE, Save the
Children, and World Vision. Enterprise Development and Microfinance 28 (1–2),
44–62.
Ochola, W.O., Sanginga, P.C., Bekalo, I., 2010. Managing Natural Resources for
Development in Africa. IDRC (538 pp).
Okonya, J.S., Kroschel, J., 2014. Gender differences in access and use of selected pro-
ductive resources among sweet potato farmers in Uganda. Agriculture & Food
Security 3 (1).
Polar, V., Babini, C., Flores, P., 2015. Technology for Men and Women: Recommendations
to Reinforce Gender Mainstreaming in Agricultural Technology Innovation Processes
for Food Security. AGRIDEL SRL, International Potato Center, La Paz - Bolivia 44 pp.
Popescu, A.D., Borca, C., Fistis, G., Draghici, A., 2014. Cultural diversity and differences
in cross-cultural project teams. Procedia Technology 16, 525–531.
Prowse, M., 2010. Integrating Reflexivity into Livelihoods Research. Progress in
Development Studies 10 (3), 211–231.
Reed, M.S., 2008. Stakeholder participation for environmental management: a literature
review. Biol. Conserv. 141 (10), 2417–2431. http://dx.doi.org/10.1016/j.biocon.
2008.07.014.
Riisgaard, L., Bolwig, S., Ponte, S., du Toit, A., Halberg, N., Matose, F., 2010. Integrating
poverty and environmental concerns into value-chain analysis: a strategic framework
and practical guide. Development Policy Review 28, 195–216.
Roy, S.K., Fuchs, G.J., Mahmud, Z., Ara, G., Islam, S., 2005. Intensive nutrition education
with or without supplementary feeding improves the nutritional status of moderately-
malnourished children in Bangaldesh. J. Health Popul. Nutr. 23 (4), 320–330.
Schindler, J., Graef, F., König, H., Mchau, D., Saidia, P., Sieber, S., 2016a. Sustainability
impact assessment of agricultural upgrading strategies to improve food security.
Environ. Impact Assess. Rev. 60, 52–63.
Schindler, J., Graef, F., König, H.J., Mchau, D., 2016b. Developing community-based food
security criteria in rural Tanzania. Food Security. http://dx.doi.org/10.1007/s12571-
016-0627-1.
Schindler, J., Graef, F., König, H., 2016c. Participatory impact assessment: Bridging the
gap between scientists' theory and farmers' practice. Agric. Syst. 148, 38–43.
Silvestrini, S., 2011. Ex-Ante-Evaluation. Ein Planungsansatz für die
Entwicklungszusammenarbeit, 10 ed. Waxmann Verlag GmbH, Göttingen.
Stauffacher, M., Flüeler, T., Krütli, P., et al., 2008. Analytic and dynamic approach to
collaboration: a transdisciplinary case study on sustainable landscape development in
a Swiss prealpine region. Syst. Pract. Action Res. 21 (6), 409–422.
Thow, M.E., Summers, M.J., Summers, J.J., Saunders, N.L., Vickers, J.C., 2017. Variations
in the APOE allele or BDNF Val66Met polymorphism are not associated with changes
in cognitive function following a tertiary education intervention in older adults: the
Tasmanian Healthy Brain Project. Neurobiol. Aging 55, 175–176.
Tittonell, P., van Wijk, M.T., Rufino, M.C., Vrugt, J.A., Giller, K.E., 2007. Analysing trade-
F. Graef et al.
offs in resource and labour allocation by smallholder farmers using inverse modelling
techniques: a case-study from Kakamega district, western Kenya. Agric. Syst. 95
(1–3), 76–95.
Trans-SEC, 2016. http://project2.zalf.de/trans-sec/public/factsheet (accessed 15th Dec.
2016).
Tsikata, D., Yaro, J.A., 2014. When a good business model is not enough: land transac-
tions and gendered livelihood prospects in rural Ghana. Fem. Econ. 20 (1), 202–226.
89
Environmental Impact Assessment Review 68 (2018) 81–89
USAID, 2008. Preliminary rural livelihood zoning: Tanzania. In: A Special Report by the
Famine Early Warning System Network (FEWS NET). USAID, Dar es Salaam.
White, G., Ruther, M., Kahn, J., 2016. Educational inequality in India: an analysis of
gender differences in reading and mathematics. In: IHDS Working Paper 2016-2.
Zimmerer, K.S., Carney, J.A., Vanek, S.J., 2015. Sustainable smallholder intensification in
global change? Pivotal spatial interactions, gendered livelihoods, and agrobiodi-
versity. Curr. Opin. Environ. Sustain. 14 (0), 49–60.